Real-time control of a robot arm using simultaneously recorded neurons in the motor cortex

Area of Science:

• Neuroscience
• Robotics
• Biomedical Engineering

Background:

• Real-time control of external devices using neural signals is a key goal in neuroprosthetics.
• Motor cortex activity can potentially be decoded to infer movement intentions.

Purpose of the Study:

• To investigate the feasibility of using simultaneously recorded motor cortex neurons for real-time robot arm control in rats.
• To assess the efficacy of a neurorobotic system in enabling device manipulation based on neural population activity.

Main Methods:

• Rats were trained to operate a robot arm using a lever for water reward.
• Multineuron signals from the motor cortex were recorded and transformed into 'neuronal population functions' using mathematical models, including neural networks.
• These functions were converted into real-time electronic signals to control the robot arm in a neurorobotic mode.

Main Results:

• Neuronal population functions accurately predicted lever trajectory.
• Four out of six rats successfully controlled the robot arm using brain-derived signals in neurorobotic mode.
• With continued training in neurorobotic mode, animals showed reduced or ceased lever movement, indicating direct neural control.

Conclusions:

• Simultaneously recorded motor cortex neurons can be effectively decoded for real-time control of external devices.
• This study presents a viable neurorobotic approach for movement restoration, offering hope for paralysis patients.
• The findings highlight the potential of brain-computer interfaces for functional recovery after neurological injury.